A column about design process, design policy, design and politics.

“ …towns and buildings will not be able to become alive, unless they are made by all the people in society, and unless these people share a common pattern language, within which to make these buildings, and unless this common pattern language is alive itself…” Christopher Alexander et al, A Pattern Language, OUP, 1977.

Pattern books for medium density housing won’t work

The NSW government has suggested that the approvals process for medium density housing schemes could be improved if there was a ‘pattern book’ of accepted designs. They’ve even foreshadowed an international design competition. But I think the idea is doomed.

There’s a huge variety of possible sites, and finding a pathway through the tortuous current controls to establish ‘standardised designs’ for those sites leads to the conclusion that a universal pattern book requires an almost infinite number of possible designs. Which defeats the purpose.

This was our discovery when we explored the idea of developing standard designs to fit a variety of sites for duplexes and terraces some time ago. Although the ideas are strong, they had limited value, and the big takeaway was that it’s a fool’s errand.

Before I show our proposals, let me try to explain the many problems with pattern books, not only how improbable they might be, moreover how the very idea insults the core skills of architects.

Variety of sites

Any site for a medium density housing scheme development site has at least 5 variables: it’s topography in area, shape and slope; it’s orientation in aspect + prospect; it’s landscape in geology, trees and vegetation; it’s context in respect of the street, vehicle access and neighbours; and the available services of electricity, water, sewer and stormwater.

Mathematically that means there are thousands of variations. Effectively it means every site is unique. Whilst it may be possible to imagine ‘idealised sites’, long experiences with other forms of standardisation, such as single project homes, shows just how difficult that can be. A discussion with any builder at ‘Homeworld’ will show how much adaptation even for a solo house needs.

Variety of controls

Design controls for housing are in both local Council and State codes. Sometimes contradictory, the confusion often makes for cumbersome complexity. Councils set the zonings, which establishes the type of housing permitted: semidetached / townhouses / boarding houses / low, high or shop-top apartments.

They also set the ‘envelope controls’ of FSR, height and setbacks for a site, dictating the size and form of a building. The 50+ councils in greater Sydney jealously guard their right to set their own rules according to ‘local conditions’, so there is little consistency in overall parameters, despite common problems.

On the other hand, the State Environment Planning Policies, or SEPPs, set out the detailed design requirements for the housing types. Inconsistently, the SEPP for apartments is far more demanding than the one for boarding houses, even though the latter may be just as big, or bigger, as we discussed last week.

But modest studio apartments of about 30 sqm, which are much in demand in the current housing crisis, cannot be built. The banks won’t fund them for purchase, the government won’t legislate for sale or rent. So, all those inner-city small units from sixty and seventy years ago are now commanding huge values. Will the new pattern book defy their own limited rules, and recognise them?

Codifying these diverse controls into simplified standardised designs or ‘patterns’ that can be used universally would be a herculean task. One council allows ten terraces in a row, another only four. Some require maximum carparking, others none. One allows basements, another doesn’t. Some allow stormwater to be pumped up, others require easements. It’s a vipers nest as any architect can tell you.

And all that before we get into the vexed issue of aesthetics and ‘design excellence’, a gold mine for NIMBY objectors. Gaining agreement from councils and the community for even some simple syntax rules would be a remarkable achievement.

Will the State legislate the ‘pattern designs’, as they did for the Complying Development Code, to override the councils and community? To which Councils responded by invoking huge swathes of heritage areas and other overlays, to thwart those designs in their LGA. An ugly stand-off.

An insult to architects and the suburbs

Given those variables of sites and controls, the combinations are virtually infinite. The variety of sites is so vast, not even an expansive catalog of patterns could satisfy all the possible permutations. Every site needs a unique solution.

Which begs the question – isn’t that what architects are good at? Creating an individual design tailored to a particular site, meeting all the rules, and solving the unique problems.

Trying to develop a taxonomy of typologies that will make approval quicker has two dreadful outcomes:

First, it's an insult to the profession of architecture to suggest you can codify and dispense with the key skillset architects bring to a project; second, if a limited pattern book was introduced, it will create a plethora of cookie-cutter solutions that will degrade our suburbs and leave councils and their communities infuriated.

To be frank: the whole idea won’t work, it won’t fix the assessment time issue, and it insults the very creatives whom it was originally intended to help. Here’s some evidence.

Two attempts, two failures

Here are two of four schemes we did for the ‘Missing Middle Competition’ in 2016-7. They weren’t winners, not by a long chalk.

The first ‘pattern book’ codifies duplexes, so they could always have north sun into the living areas and private open space. Nice idea, but what’s ‘north’, how many orientations are there? What about street address? What happens when you put the living and external areas on the second level for the duplex on the south?

We developed three very different designs and a wagon-wheel of possibilities (nominative determinism). They are far more inventive, and better, than the vast range of CDC duplexes that have proliferated in the 7 years since. But the ‘pattern’ was far too radical in approach; it wasn’t even shortlisted at the time.

The second ‘pattern book’ looks at small terraces, on a single house block, and takes the primacy of orientation versus all the usual conceits to allow sun to six units, but also solar access to the surrounding properties. Given few sites are cardinal N-S or EW, we developed some angled floor plans to get the north sun, and nick-named them the ‘cranky terraces’.

Bookends

A far better way to codify design was set out in this book, A Pattern Language, in 1977 by Christopher Alexander et al, OUP, 1977. Much out of fashion now, this book showed how multiple patterns can be used to build an individual work. It is how patterns are meant to work.

Next week: envelope controls as an alternative

Next week, I will put forward and alternative. Instead of trying to codify the solutions, why not turn it on its head and codify the problem? By which I mean having far more precise design controls on each site, such that a design that complies within those controls completely is approved.

But how to arrange those controls so they are not too prescriptive? The answer is to use modern 3D technology to do away with the clunky ideas of FSR, height and setbacks. To have a unique ‘spatial envelope’, within which a design can be developed, knowing it is automatically approved. More next week.

Tone Wheeler is an architect /adjunct prof UNSW / president AAA.

The views expressed are his.

These Design Notes are Tone on Tuesday #199, week 11/2024.

Past Tone on Tuesday columns can be found here

Past A&D Another Thing columns can be found here

You can contact TW at [email protected]